I made that video and here's what I'll say:
There is no constant number to add to create an equivalence in load times between versions. Hell, even different models of Wii's load faster than others. It also depends if you have already loaded the track prior to starting (say you play Luigi's Circuit once before starting a run). Also note that different hard drives on different computers all load the game at different speeds on Dolphin as well.
What's more likely to happen is we time using the in-game timer instead of RTA. Even then, I'm not sure I would allow Dolphin. The Ubershaders update has improved stuttering on MKWii, but I still don't think performance is perfect.
Thanks for that video, must've been hard to make.
"What's more likely to happen is we time using the in-game timer instead of RTA."
And essentially mandate spreadsheet programs to add up times? :/
"improved stuttering"
Does that make runs 'faster'?
"There is no constant number to add to create an equivalence in load times between versions. "
True, but say the difference between loading times is 4-6 seconds. I still think it's perhaps useful to note that to some degree, even if you cant find the specific value between the 4 and 6 second mark. This wide gap, although flawed, is still useful IMO, because it still allows viewers to make a generalized difference between two runs without watching them in full.
For example, PLAYER wants to watch a fast speedrun around the 5th place mark as PLAYER believes he wants to attain a 5th place position. 5, 6, and 7th places are all tied, and on different console types. Using these generalised time gaps between loading differences, PLAYER can find out the best run to watch for tips etc for his time range as the one on console that has the slowest loading time would produce the fastest, most skilful speedrun. So he watched that one and learns more and gets a more realistic sense of the pace he needs to be at. Not an exact pace, but fairly accurate pace. Useful!
Now, say one console saves roughly 28 to 32 seconds. A player uses this console, and an equally skilled player on another console attains a time 25 seconds slower. It's still obvious that the player who got 25 seconds slower is still ACTUALLY THE MORE SKILLFUL RUN, because the time range is 28 to 32 seconds faster. 25 is much lower than the most severe time range thing of 28 seconds. 25-LOWER-THAN-28. xd
tl;dr console differences don't have to be exact and are still useful and should probably be noted as they are still able to provide useful time range differences. However it's not exact enough to break leaderboard ties down to the millisecond or anything like that, but it still provides a useful time range comparison.
If you have a fear of spreadsheets, LiveSplit has functionality for in-game time input. Other games like Sonic use this (though for reasons unrelated to version differences).
My issue with stuttering in Dolphin is that it's not 100% at parity with console. Generally, games that allow emulator runs have identical performance between console and emulator. The top Wii games I could find that allow Dolphin besides Project M (which isn't even an official game) are Kirby's Return to Dream Land and Super Mario Sluggers which have one emulator run submitted each. The majority of Wii games do not allow Dolphin runs, and in my opinion Dolphin is not at a state where speedruns on it can be comparable to console.
If you want to use estimated load time differences to compare runs across platforms, go ahead. We can't change the leaderboard with these estimates, but I agree that you should consider loads when looking for the fastest run to learn from.
Having to type out your entire 10 character time format thing after each race while in the middle of a speedrun sounds really tedius, distracting, and difficult, and I suspect you have to do it that way with Livesplit.
"We can't change the leaderboard with these estimates"
you're pretty much right, but you could use those esitmates to find the highest ranges of difference. ie "itll NEVER be slower than 32 sec and it'' NEVER be faster than 28 sec" so you could use that gap to have a leaderboard with gaps but i dont think it's feasible, but bloody hell it'd be accurate
and i was basically asking what stuttering is in dolphin
"games that allow emulator runs have identical performance between console and emulator. "
so why would dolphin be banned if the times are the same??
^Yeah but...with live recordings now being mandatory for the top runs, wouldn't savestates/cheated runs be impossible?
No, you can run a script or make inputs yourself to play out a whole run and stream it on Twitch or whatever. It actually isn't that hard might I add.
Savestates wouldn't work for that - even if you scripted something to make savestates, you wouldn't be able to use the savestates - it'd be in video of you using them!
what kind of scripts/inoputs are you talking about that would go undetected on stream? asking because i dont see how thats possible,
You can easily load one savestate that shows before you start the stream and then it'll play out a few attempts and then a full run. I don't know the most about emulators, I just know that it is possible to cheat on stream. RNG manipulation via emulator is also a thing.
Does stuttering makes runs longer? Or do you just see less frames but it's still played in realtime?
I don't remember what 'Speed up disc transfer rate' was set to. Do keep in mind that one of the Dolphin tests was done on a physical disc drive with a burned backup disc.
You can submit a run to the leaderboard with in-game time, but it must also include real time as the leaderboards are not fully migrated to in-game time.
